Prescription drugs accounted for 11% of Medicare spending in 2010, but are expected to account for 20% of this budget by 2020. The purpose of this subproject is to advance our understanding of the use and efficiency of prescription drugs among Medicare beneficiaries. First, we characterize the use of pharmaceutical drugs in the U.S. Medicare population at the level of the region or physician-hospital network (PHN) into four categories depending on efficiency or effectiveness - that is, value per dollar spent. These range from the first category, highly effective, to the fourth category, drugs which could lead to adverse effects on health. In this and subsequent analyses, we will consider factors that may affect the efficiency of pharmaceutical use, such as care fragmentation, academic affiliation, health information technology adoption, market structure, and for some physicians, board scores from the American Board of Internal Medicine examinations. These measures can also be used across other subproject analyses, for example for PHNs in subproject 1 and regions in subproject 4. Second, we focus on pharmaceutical treatments for specific cohorts: myocardial infarction survivors, hip fracture survivors, and incident cases of chronic pulmonary disease. This focus allows us to adjust more easily for patient level factors and identify clearly effective treatments (e.g., clopidogrel following a drug-eluting stent). In later years, data will permit assessment of the impact of payment structure such as shared savings and partial capitation on prescription use measures. Third, we use the cohorts developed in the previous subsection to measure substitution effects, the extent to which spending wisely on highly effective pharmaceuticals may have positive effects on downstream healthcare costs, and conversely for those drugs with little known value (or even negative value). To adjust for differences across areas in underlying risk, we will draw on Project 5's risk adjustment measures. Finally, we describe differences in prescription drug use in population subgroups. Early work shows substantial differences in use of some effective medications among Medicare subgroups (e.g. Black, older, poor) compared to the population overall. In addition, we intend to study the under-65 Medicare population (largely those receiving Social Security Disability Insurance) where pharmaceutical treatment for mental illness is likely to be common.